Phasecraft reports its AI agents achieved up to 23% lower Pauli weight in fermion-to-qubit encodings, surpassing the performance of an optimizer and demonstrating a quantifiable improvement for quantum simulation. The company’s system employs AI agents co-created with Phasecraft researchers, then “frozen” after validation to ensure verifiable and repeatable results. The researchers state that every solution is a program that can be read, and the paper identifies the idea behind each improvement. By reducing the need for SWAP gates by as much as 90%, this approach directly addresses a key obstacle to scaling quantum computations.
AI Agents and Researchers Co-Develop Quantum Verifiers
Phasecraft’s system achieves up to 90% fewer SWAP gates in qubit layouts, a reduction in a computationally expensive operation that limits quantum circuit scale. This improvement stems from a novel approach where AI agents collaborate with Phasecraft researchers, who ultimately validate them. Each agent is co-created with a researcher who certifies it through a series of tests before the AI’s design is “frozen” for solution submission.
This process ensures verifiable and repeatable results, avoiding the opacity of fully autonomous AI systems. The company tested these AI-driven optimizers on three quantum simulation problems, consistently exceeding the performance of existing methods, Phasecraft says. The approach also yielded up to 73% fewer CNOT gates during circuit synthesis, further streamlining quantum computations. Phasecraft believes this methodology extends beyond the tested problems, offering a broadly applicable framework for optimizing various quantum computing tasks.
See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
